POSNER, Circuit Judge.
We decided to hear this case en banc, pursuant to Circuit Rule 40(f), in order to decide whether violation of the federal “kingpin” statute, 21 U.S.C. § 848 (Continuing Criminal Enterprises, Title II, § 408, of the Organized Crime Control Act of 1970), is “an offense against the United States” within the meaning of the federal aider and abettor statute, 18 U.S.C. § 2(a). That statute provides: “Whoever commits an offense against the United States or aids, abets, counsels, commands, induces or procures its commission, is punishable as a principal.” A panel of this court had answered the question “yes” in United States v. Ambrose, 740 F.2d 505, 507-08 (7th Cir.1984), followed by a district judge in United States v. Vasta, 649 F.Supp. 974, 982 (S.D.N.Y.1986). Subsequently the Second Circuit, emphasizing legislative history not discussed in Ambrose, answered “no” in United States v. Amen, 831 F.2d 373, 381-82 (2d Cir.1987), followed in United States v. Benevento, 836 F.2d 60, 71-72 (2d Cir.1987). Oddly — when one considers that the kingpin statute has been on the books for almost twenty years — no other court has decided whether there can be liability for aiding and abetting a drug kingpin. The question is alluded to in United States v. O’Connell, 841 F.2d 1408, 1425 n. 9 (8th Cir.1988), but not decided.
A court of appeals has a responsibility to reexamine its decisions in light of new arguments, new evidence, new experience, especially when by doing so it may be able to eliminate a conflict between circuits and thereby lighten the Supreme Court’s burden of resolving such conflicts. In that spirit we have undertaken to reexamine Ambrose, but having done so we adhere to our view that there is aider and abettor liability for assisting a kingpin.
The kingpin statute imposes heavy penalties for the commission of a felony narcotics violation as part of a continuing series of violations from which the perpetrator obtains substantial income or resources and which he conducts in concert with five or more persons with respect to whom he “occupies a position of organizer, a supervisory position, or any other position of management.” 21 U.S.C. § 848(b)(2). (This section was renumbered to (d)(2)(A) pursuant to amendments made to the kingpin statute in 1986, but Pino-Perez was convicted under the original statute.) The minimum penalty is ten years in prison and the maximum is life in prison plus a $100,-000 fine (raised to $2 million in 1986). 21 U.S.C. § 848(a)(1). There is no parole, see § 848(c), although time off for good behavior and work in an industry or camp can cut the kingpin’s sentence by more than a third. See 18 U.S.C. §§ 4161, 4162. The new federal sentencing statute enacted in 1986 abolishes parole for all federal crimes committed after its effective date (and time off for good behavior has also been trimmed), but when the kingpin statute was originally enacted its provision disallowing parole was unusual. Suspension and probation are also disallowed. See 21 U.S.C. § 848(c). Congress imposed a stiff minimum mandatory prison sentence — and meant it.
As the government has rightly conceded in these cases, the persons supervised by the kingpin cannot be punished as aiders and abettors. See United States v. Ambrose, supra, 740 F.2d at 507-08; United States v. Amen, supra, 831 F.2d at 381-82. When a “crime is so defined that participation by another is necessary to its commission,” that other participant is not an aider and abettor. United States v. Southard, 700 F.2d 1, 20 (1st Cir.1983). “[B]y specifying the kind of individual who is to be found guilty when participating in a transaction necessarily involving one or more other persons, [the legislature] must not have intended to include the participation by others in the offense as a crime. This exception applies even though the statute was not intended to protect the other participants.” Id. The exception covers persons whom a kingpin supervises. [1232] If they were chargeable as his aiders and abettors, the purpose of the kingpin statute —to punish the kingpin more severely than other drug offenders — would be thwarted. Since the lowliest mixer, if punishable as an aider and abettor, would be subject to the same ten-year minimum as the kingpin himself, there would be no incremental deterrence of the kingpin in cases where the sentencing judge thought the statutory minimum adequate for the kingpin. Cf. United States v. Farrar, 281 U.S. 624, 50 S.Ct. 425, 74 L.Ed. 1078 (1930).
Southard describes two other exceptions to aider and abettor liability (see 700 F.2d at 19-20). The first concerns the victim of the crime. Even if, as in such crimes as extortion, blackmail, and bribery, his conduct significantly assisted in the commission of the crime, he cannot be charged as an aider and abettor. The second exception concerns members of a group that the criminal statute seeks to protect: a woman who is transported willingly across state lines for the purpose of prostitution cannot be charged as an aider and abettor of the transporter’s Mann Act violation. Gebardi v. United States, 287 U.S. 112, 123, 53 S.Ct. 35, 38, 77 L.Ed. 206 (1932).
Persons who assist a kingpin but are not supervised, managed, or organized by him do not fit any of these three exceptions, and we are reluctant to create a fourth. Cf. United States v. Santore, 290 F.2d 51, 76-78 (2d Cir.1960) (en banc). In Ambrose, the kingpin’s aiders and abettors were police officers who protected the kingpin’s operation. The role of the aider and abettor in Amen, Paradiso, is unclear from the Second Circuit’s opinion; since the Second Circuit rejected aider and abettor liability, it had no reason to particularize Paradiso’s activities. The government’s brief in Amen describes Paradiso as “Abbamonte’s [the kingpin’s] trusted friend who assisted him in directing his distribution network from prison by arranging to punish one of his workers and by arranging for the acquisition of heroin. Without in any sense being employed or supervised by Abbamonte, Paradiso provided valuable assistance to the head of the network and his underlings.” Brief for the United States of America in Nos. 87-1028 et al. (United States v. Amen), at p. 42. And United States v. Vasta, supra, a decision on pretrial motions in the same prosecution, reports the government’s contention that the assistance of Paradiso and another alleged aider and abettor, Squitieri, “was of critical importance in keeping Abba-monte’s operation alive.” 649 F.Supp. at 982. Ernest A. Benevento, the person charged with aiding and abetting a kingpin in United States v. Benevento, supra, “engaged in managerial functions throughout J.E.M.’s drug operations [J.E.M. was an international criminal enterprise engaged in the manufacture and distribution of heroin] and provided substantial assistance to Ernesto J. Benevento [the kingpin. Ernest] ... provided J.E.M. with the use of his Arizona home for the drug manufacturing laboratory, maintained detailed financial records for the enterprise on his computer at home, contributed substantial amounts of capital to fund the second drug venture and attempted to smuggle currency out of the United States for J.E.M.” 836 F.2d at 71. The aider and abettor in the present case, Pino-Perez, supplied cocaine to a drug ring in southern Wisconsin headed by Harold Nichols. Pino-Perez was the supplier to Nichols’s enterprise, and the quantities supplied were impressive. Pino-Perez once received $30,000 in cash for two kilograms of cocaine that he sold to the Nichols enterprise. He sold the enterprise as much as five kilograms at a time, and planned a final sale of 20 kilograms to it. His sales were frequent, and they continued for a long time. He claimed to know all or at least most of the big cocaine dealers in the country. This may have been braggadoccio, but there is no question that he was a big dealer — much bigger than Nichols, the kingpin.
The judge sentenced Pino-Perez to 40 years in prison under the kingpin statute for aiding and abetting Nichols’s enterprise. Besides asking us to overrule Amb-rose, Pino-Perez contends that the trial demonstrated either a fatal variance between the indictment and the proof or a “constructive” amendment of the indict[1233] ment. This contention has no possible merit and requires no discussion. See United States v. Leibowitz, 857 F.2d 378, 378-79 (7th Cir.1988). His remaining contentions are equally meritless, so we may turn to the issue of aider and abettor liability.
What is now section 2(a) of the federal criminal code dates back to 1909. See Act of March 4, 1909, ch. 321, § 332, 35 Stat. 1152; Standefer v. United States, 447 U.S. 10, 15-20, 100 S.Ct. 1999, 2003-06, 64 L.Ed. 2d 689 (1980). Ever since then, every time Congress has passed a new criminal statute the aider and abettor provision has automatically kicked in and made the aiders and abettors of violations of the new statute punishable as principals. “With the enactment of that section [the 1909 predecessor to section 2(a) ], all participants in conduct violating a federal statute are ‘principals.’ ” Id. at 20, 100 S.Ct. at 2006 (dictum). “The aiding and abetting provision [section 2(a) ] ... is applicable to the entire criminal code.” United States v. Jones, 678 F.2d 102, 105 (9th Cir.1982). To the same effect, see, e.g., United States v. Sopczak, 742 F.2d 1119, 1121 (8th Cir.1984) (per curiam); United States v. Rector, 538 F.2d 223 (8th Cir.1976) (per curiam). No cases other than Amen and Benevento hold section 2(a) totally inapplicable to a federal criminal statute. Not every accomplice in the commission of a federal criminal offense is an aider and abettor, but until Amen every aider and abettor of a federal criminal offense had been thought punishable for the offense as a principal by virtue of section 2(a). The prostitute transported across state lines is not an aider and abettor of a Mann Act violation, but a supplier of prostitutes to the transporter may be (depending on what he knows — more on this shortly), and he is therefore punishable by virtue of section 2(a) as a principal violator of the Mann Act. United States v. Simmons, 610 F.Supp. 295, 304 (M.D.Tenn.1984). The kingpin’s employees are not aiders and abettors, but — one might have thought— the supplier of the kingpin’s drugs may be.
In reaching a contrary conclusion, Amen relied primarily on the legislative history of the kingpin statute. Early versions of the bill that became section 848 were sentence-enhancement provisions under which if a defendant was convicted of a drug offense (such as manufacture, distribution, or possession with intent to distribute, see 18 U.S.C. § 841(a)(1)), the prosecutor would have been allowed to argue to the sentencing judge that the defendant should be punished more heavily than usual because he was a drug kingpin. See H.R.Rep. No. 1444, 91st Cong., 2d Sess. 80-84 (1970) U.S. Code Cong. & Admin.News 1970 pp. 4566, 4648, 4652 (additional views of four Congressmen). If the sentence-enhancement format had been retained in section 848 as enacted, there would have been no aider and abettor liability for assisting a kingpin, because there would have been no kingpin offense; there would just have been kingpin offenders against other criminal statutes. But it was not retained. In response to objections that in a sentencing hearing the defendant would not be able to cross-examine those who had accused him of being a kingpin, the bill was amended to make “engagement in a continuing criminal enterprise a new and distinct offense with all its elements triable in court.” Id. at 84. From this history the Second Circuit concluded in Amen that “while the legislative history makes no mention of aiders and abettors, it makes it clear that the purpose of making CCE [Continuing Criminal Enterprises — the official name of the kingpin statute] a new offense rather than leaving it as sentence enhancement was not to catch in the CCE net those who aided and abetted the supervisors’ activities, but to correct its possible constitutional defects by making the elements of the CCE triable before a jury.” 831 F.2d at 382. To this we cannot say “amen.” True, it was not Congress’s purpose in making the operation of a continuing criminal enterprise a separate offense to bring section 2(a) into play. But such is never Congress’s purpose in creating a new offense. Congress doesn’t have to think about aider and abettor liability when it passes a new criminal statute, because section 2(a) attaches automatically. The question is not whether section 2(a) is applicable — it always is. The [1234] question is whether a given accomplice is an aider and abettor. The statute, here the kingpin statute, is used to determine who counts as an aider and abettor, for the term is not defined by section 2(a). But once that determination is made, liability is automatic by virtue of section 2(a).
It would introduce great uncertainty into federal criminal law if the liability of a conceded aider and abettor depended on the results of an inquiry into Congress’s intent concerning such liability in creating the offense that the defendant aided and abetted. Yet that is the inquiry required by Amen. The passage we have quoted could even be interpreted to mean that unless a specific intent to punish aiders and abettors appears in the legislative history of a criminal statute, section 2(a) does not apply to that statute; aiding and abetting violations of the statute is not a crime. That approach would essentially abolish federal aider and abettor liability. A more modest version of the approach would require the kind of inquiry that the Supreme Court uses to decide whether a regulatory statute that is silent on private rights of action can be enforced by them: it asks “whether Congress, either expressly or by implication, intended to create a private right of action.” Touche Ross & Co. v. Redington, 442 U.S. 560, 576, 99 S.Ct. 2479, 2489, 61 L.Ed.2d 82 (1979). Until Amen the presumption with regard to aiding and abetting had been different. There had to be “ ‘an affirmative legislative policy’ to create an exemption from the ordinary rules of accessorial liability.” United States v. Falletta, 523 F.2d 1198, 1200 (5th Cir.1975), quoting Gebardi v. United States, supra, 287 U.S. at 123, 53 S.Ct. at 38. Doubt about Congress’s intentions was resolved in favor of aider and abettor liability.
The opinion in Amen gives the impression that the conversion of the kingpin statute from a sentence-enhancement provision to a provision creating a new and distinct offense was an unconsidered last-minute switch. That is not correct. The Dingell Amendment (which brought about the switch) was debated extensively; what is more, the sentence-enhancement provision was ultimately restored to another provision of Title II of the omnibus act. See 21 U.S.C. § 849 (dangerous special drug offender sentencing). The full story is told in Garrett v. United States, 471 U.S. 773, 780-85, 105 S.Ct. 2407, 2412-15, 85 L.Ed.2d 764 (1985). The new and distinct offense created by section 848 was more fully debated than most federal criminal statutes, and there is no more reason to infer from its legislative history an intent to preclude aider and abettor liability than there would be to draw such an inference from the legislative history of any other federal criminal statute. The three exceptions recognized in Southard exhaust the cases in which an inference can confidently be drawn that Congress in enacting a criminal statute meant to protect a class of accomplices from being charged as aiders and abettors.
Besides legislative history, Amen emphasizes practical objections to aider and abettor liability in kingpin cases. “How does one determine whether a person is an employee [of the kingpin, and hence not an aider and abettor] or third party? What of the businessman who leases a boat to a CCE engaged in importation? What about the kingpin’s bodyguard? Or his lawyer?” 831 F.2d at 382. These are legitimate rhetorical questions, and they can be multiplied. If there is aider and abettor liability in kingpin cases, could not assistants outside the organization be deemed aiders and abettors and be punished more severely than important operatives within the organization? And would not assistants with roles in two or more organizations be simultaneously covered and excluded? Suppose the main supplier to a distribution network in Wisconsin was the henchman of a larger network in Colombia that had subsidiaries in Florida. If the government prosecuted the Colombia or Florida operations, their operatives could not be convicted as aiders and abettors, because they would be subordinates; but if the government prosecuted only the Wisconsin operation, a member of the Florida organization could be convicted because he was not a subordinate of the Wisconsin kingpin. Indeed, someone charged as an aider and [1235] abettor of a kingpin might feel impelled to show that he was really one of the kingpin’s employees — which would get him out of the section 848 frying pan but land him in a section 841 fire. And the supplier of a small quantity of drugs to a kingpin — a quantity too small to permit severe punishment under section 841 — would, if liable as an aider and abettor of the kingpin, be severely punishable under section 848. See 21 U.S.C. § 841(b).
These rhetorical questions and classroom-type hypotheticals sound more ominous than they are. (The absence of any litigation over the issue of which accomplices of a drug kingpin are aiders and abettors provides some reassurance on this score.) Let us consider each of them. The lessor of the boat would be an aider and abettor of the kingpin — provided he met the requirements for an aider and abettor. This is an important qualification and it undercuts Amen’s position. We and other courts have endorsed Judge Learned Hand’s definition of aiding and abetting, which requires that the alleged aider and abettor “in some sort associate himself with the venture, that he participate in it as in something that he wishes to bring about, that he seek by his action to make it succeed.” United States v. Peoni, 100 F.2d 401, 402 (2d Cir.1938); see United States v. Fountain, 768 F.2d 790, 797-98 (7th Cir.1985); United States v. Ambrose, supra, 740 F.2d at 509, and cases cited there; also Bosco v. Serhant, 836 F.2d 271, 279 (7th Cir.1987); United States v. Stanchich, 550 F.2d 1294, 1300 (2d Cir.1977). The venture in a kingpin case is, of course, the continuing criminal enterprise. It is that enterprise which the boat lessor in Amen’s example would have to be found to have associated with, participated in, and sought by his action to make succeed, in order to be punishable as an aider and abettor. The mere fact of leasing a boat to a person known to be a drug trafficker would not be enough to make him guilty of aiding and abetting a drug kingpin. And so with the last hypothetical as well. One who sells a small — or for that matter a large-quantity of drugs to a kingpin is not by virtue of the sale alone an aider and abettor. It depends on what he knows and what he wants: Does he want the kingpin’s enterprise to succeed or is the kingpin just another customer? If he does want the enterprise to succeed, there is no anomaly in holding him liable as an aider and abettor. Instead of buying from one large supplier (Pino-Per-ez), kingpin Nichols might have decided to buy from a host of small suppliers. If they knew he was a kingpin, associated themselves with his enterprise, participated in it as something they wanted to see succeed, and by their action sought to make it succeed, then they would indeed be punishable as aiders and abettors — and rightly so. The application of the Peoni test will sometimes require difficult factual appraisals, but the paucity of reported cases suggests that the factfinding task is a manageable one.
The bodyguard mentioned in Amen would presumably be a person supervised by the kingpin, hence not an aider and abettor. A lawyer is of course not an aider and abettor of his clients, whether or not he fits the Peoni definition; this is merely a commentary on the limitations of definition. A drug dealer could be both the underling of one kingpin and the aider and abettor of another. Cf. United States v. Bond, 847 F.2d 1233, 1236-37 (7th Cir.1988). So much the worse for him; he need not be the top man to be convicted of the kingpin offense. We so held in United States v. Moya-Gomez, 860 F.2d 706, 745-49 (7th Cir.1988); see also United States v. Becton, 751 F.2d 250, 254-55 (8th Cir.1984), and other decisions cited in Moya-Gomez, supra, at 746. For all we know, Pino-Perez is the employee of some superkingpin. He could not be convicted of aiding and abetting that kingpin, but so what? Granted, it may sometimes be to the advantage of one accused of being a kingpin’s aider and abettor to inculpate himself as an underling; but nothing is more common than for a criminal defendant to seek to exculpate himself of a major offense by accusing himself of a lesser one.
Finally, it is unlikely that important operatives of a drug kingpin would be punished less severely than less important aiders [1236] and abettors. Kingpins are not limited to heads of organizations; a kingpin is anyone who has supervisory or managerial responsibility with regard to five or more members of the organization. A large organization will contain more than one person with such supervisory responsibilities. In United States v. Rosenthal, 793 F.2d 1214, 1225-26 (11th Cir.1986), both the boss of the enterprise and one of his foremen were convicted of the kingpin offense. Even Nichols’s relatively modest enterprise had two “managers,” besides himself. Moreover, drug underlings are subject to punishment under section 841, which has its own severe minimum penalties. And, on the other hand, not every accessory is an aider and abettor; as Judge Hand’s definition and the history of section 2(a) show, an accessory after the fact is not an aider and abettor. See, e.g., Bollenbach v. United States, 326 U.S. 607, 611, 66 S.Ct. 402, 404-05, 90 L.Ed. 350 (1946); United States v. Varelli, 407 F.2d 735, 749 (7th Cir.1969); United States v. Covelli, 738 F.2d 847, 860 n. 18 (7th Cir.1984); S.Rep. No. 10, 60th Cong., 1st Sess. 13 (1908). This is another reason to believe that aider and abettor liability will not create an anomalous pattern in which the more culpable are let off more lightly than the less culpable.
The remaining objection to aider and abettor liability in kingpin cases (surprisingly not mentioned in Amen even though it had greatly troubled this court in Ambrose) arises from the heavy minimum penalty that the kingpin statute imposes. The least culpable kingpin is to be punished by at least ten years in prison with no possibility of parole. Since an aider and abettor is punishable as a principal — that is, punishable under the statute creating the offense that he was found to have aided and abetted — the least culpable aider and abettor of a kingpin must likewise be punished by no less than ten years in prison with no possibility of parole. This is a harsh result, and while it did not persuade us in Ambrose that there is no aider and abettor liability in kingpin cases, it did persuade us that section 2(a) authorizes the sentencing judge to disregard the minimum penalty in section 848. 740 F.2d at 508-10. We pointed out that the concept of aiding and abetting had come into the law at a time when all felonies carried the death penalty, and had been welcomed as a device for enabling the lesser participants in a felony to be punished more leniently than by death. See 4 Blackstone, Commentaries on the Laws of England 39 (1769); Perkins, Parties to Crimes, 89 U.Pa.L.Rev. 581, 613-15 (1941). As judges acquired more and more sentencing discretion, the felt need to distinguish between principals on the one hand and aiders and abettors on the other in order to mitigate the severity of criminal punishment diminished, and the distinction was eventually discarded, as in section 2(a).
The historic purpose of aiding and abetting liability coexists uneasily with criminal offenses that carry mandatory minimum penalties, such as the kingpin statute. We said in Ambrose that “a policeman who took an isolated bribe from a kingpin but did not engage in a prolonged and systematic protection racket, as these defendants did, would still be an aider and abettor of the kingpin, and therefore under the view that the aiding and abetting statute mechanically incorporates the whole punishment schedule of section 848 would have to be sentenced to a minimum of 10 years in prison without possibility of parole. Yet if there were no minimum the judge might sentence the policeman to only two or three years in prison and the policeman would be eligible for parole after serving a third of that time.” 740 F.2d at 509. We therefore held that in sentencing a kingpin’s aider and abettor the judge was not bound by the minimum penalty provisions of the kingpin statute. Judge Wood dissented from this holding. See id. at 513-16.
Further reflection has convinced us that Judge Wood’s position is correct. Three reasons are decisive. First, section 2(a) does not contain its own schedule of punishments but instead makes the aider and abettor punishable as a principal for the offense that he aided and abetted. That is, punishment is imposed under the statute creating that offense. Here that is the kingpin statute, which imposes a minimum [1237] penalty applicable to everyone punishable under the statute — and an aider and abettor is punishable under the statute creating the offense he has aided and abetted and under no other statute. We recognize, not disapprovingly, that judges engage in a certain amount of statutory revision, especially when dealing with two statutes that were passed at different times without explicit reference to each other and that don’t fit together nicely. But simply lopping off a minimum statutory penalty for one class of violators (aiders and abettors) now strikes us, as it did Judge Wood back in 1984, as exceeding the prudent bounds of judicial creativity.
Second, while endorsing Judge Hand’s definition of aider and abettor Ambrose paid insufficient heed to the limitations built into it. “A policeman who took an isolated bribe from a kingpin” would not be likely to be convicted of aiding and abetting the kingpin’s violations — would not be likely even to know he was dealing with a kingpin. Aiding and abetting implies a fuller engagement with the kingpin’s activities. Third, and related, in no reported case has the participation of the aider and abettor been so meager relative to the kingpin’s that subjecting him to the minimum penalty in the kingpin statute would be savage or incongruous. The “Marquette 10” — the corrupt policemen who protected the kingpin in Ambrose — were not small fry, and of course neither is Pino-Perez — he was a bigger drug dealer than Nichols, the kingpin, whom he supplied. Paradiso, the aider and abettor in Amen, was not a small fry either; we know because he was sentenced not to the statutory minimum but to 40 years, of which 20 were for aiding and abetting the kingpin. Squitieri’s role was apparently a large one too, as was Ernest A. Benevento’s role in assisting the kingpin in United States v. Benevento. Experience since Ambrose teaches that the danger which concerned us in that case — the danger that ferocious mandatory minimum sentences might be meted out to minor accomplices — is, as a practical matter, insufficiently serious to warrant so athletic an exercise in interpretive creativity as attempted in that case, let alone the casting of a cloud of uncertainty over all federal aider and abettor liability, as in Amen.
Congress may want to give attention to the problem of subjecting aiders and abettors to stiff mandatory minimum criminal penalties, although we recognize both that lenity for drug offenders is not high on the current list of national priorities and that, as illustrated by United States v. Martinez-Zayas, 857 F.2d 122, 131-33 (3d Cir.1988), a drug dealer who might be convicted as a kingpin’s aider and abettor is quite likely to be punishable in any event under provisions of section 841 that themselves carry heavy mandatory minimum penalties. (Not all aiders and abettors of drug kingpins are themselves drug dealers, however, as illustrated by Ambrose.) But we think that both the aider and abettor statute and the kingpin statute mean what they say. If the mesh between them is imperfect— which it is — still the judicial cure that we attempted in Ambrose, and the even more drastic cure that our respected colleagues in the Second Circuit attempted in Amen, are, we believe, even worse than the disease.
AFFIRMED.